In vivo and in vitro escape from neutralizing antibodies 2G12, 2F5, and 4E10

Abstract

Recently, passive immunization of human immunodeficiency virus (HIV)-infected individuals with monoclonal antibodies (MAbs) 2G12, 2F5, and 4E10 provided evidence of the in vivo activity of 2G12 but raised concerns about the function of the two membrane-proximal external region (MPER)-specific MAbs (A. Trkola, H. Kuster, P. Rusert, B. Joos, M. Fischer, C. Leemann, A. Manrique, M. Huber, M. Rehr, A. Oxenius, R. Weber, G. Stiegler, B. Vcelar, H. Katinger, L. Aceto, and H. F. Gunthard, Nat. Med. 11:615-622, 2005). In the light of MPER-targeting vaccines under development, we performed an in-depth analysis of the emergence of mutations conferring resistance to these three MAbs to further elucidate their activity. Clonal analysis of the MPER of plasma virus samples derived during antibody treatment confirmed that no changes in this region had occurred in vivo. Sequence analysis of the 2G12 epitope relevant N-glycosylation sites of viruses derived from 13 patients during the trial supported the phenotypic evaluation, demonstrating that mutations in these sites are associated with resistance. In vitro selection experiments with isolates of four of these individuals corroborated the in vivo finding that virus strains rapidly escape 2G12 pressure. Notably, in vitro resistance mutations differed, in most cases, from those found in vivo. Importantly, in vitro selection with 2F5 and 4E10 demonstrated that resistance to these MAbs can be difficult to achieve and can lead to selection of variants with impaired infectivity. This remarkable vulnerability of the virus to interference within the MPER calls for a further evaluation of the safety and efficacy of MPER-targeting therapeutic and vaccination strategies.

Figures

FIG. 1.

Changes in potential glycosylation sites within C2 to V4 induced during in vivo treatment with 2G12. Samples derived at different time points from plasma bulk cultures and single clones from 13 patients (NAB01 to NAB13) following passive immunization were sequenced, and changes are summarized. The five glycosylation sites defining the 2G12 epitope are shaded in dark gray. Non-2G12-related potential glycosylation sites are shaded in light gray. X denotes a sequence mutation in the glycosylation site that lead to loss of that specific site. Glycosylation sites created through mutations during passive immunization are indicated by dashed boxes. The presence of mixed populations is indicated by boxes with diagonal divisions. The sensitivity to 2G12 (IC90) of available replication-competent viruses derived during our passive-immunization study was measured in a PBMC-based assay (52). An asterisk indicates that the isolate was sensitive to 2G12 at the IC70. Double asterisks indicate the measured IC90 for the isolate derived at week 24.

FIG. 2.

Representative profile of in vitro selection experiments performed with MAbs 2G12, 2F5, and 4E10. Individual panels depict the evolution of the resistant viral variants of isolate NAB01 for each of the three MAbs (2G12, 2F5, and 4E10) tested and all three MAbs in combination. Gray lines and symbols (gray axis) signify p24 antigen production in cultures. Shaded areas represent the antibody level present in the culture during selection (orange for 2G12, magenta for 2F5, and blue for 4E10). The same colors are used to illustrate the neutralization sensitivities (IC50, clear bars; IC70, dashed bars; IC90, filled bars) of the viruses that emerged to the respective antibodies at the indicated time points.

FIG. 3.

Overview of in vitro selection experiments with 2G12, 2F5, and 4E10. Selection experiments were conducted one to four times (indicated by roman numerals) with isolates from patients NAB01, NAB02, NAB03, and NAB08. The neutralization sensitivity (IC50, clear bars; IC70, dashed bars; IC90, filled bars) attained during each selection series is depicted. Genotypic resistance was further characterized by sequencing (see Fig. 6 and 7). When selection resulted in abortive infection, the timing of the termination of an experiment is indicated.

FIG. 4.

Overview of sequence changes within the C2-to-V4 region of isolates selected with 2G12 in vitro. The C2-to-V4 region of the 2G12-selected isolates derived from patients NAB01, NAB02, NAB03, and NAB08 are shown together with the respective mock-treated isolates and the sequence changes found in vivo during passive immunization. 2G12-relevant glycosylation sites are shaded in dark gray. Non-2G12-related potential glycosylation sites are shaded in light gray. X denotes a sequence mutation in the glycosylation site that lead to loss of that specific site. The presence of mixed populations is indicated by boxes with diagonal divisions.

FIG. 5.

Overview of sequence changes in isolates selected for resistance to 2F5 and 4E10. The epitopes of the selected isolates derived from patients NAB01, NAB02, NAB03 and, NAB08 are shown together with the respective pretreatment and mock-treated isolates. Intact epitopes for 2F5 and 4E10 are shaded in magenta and blue, respectively; mutations in the epitopes conferring resistance are in red. aa corresponds to the amino acid numbering in the HXB2 reference strain.

FIG. 6.

Neutralization sensitivities of 4E10 escape mutants. Neutralization sensitivities of Env-pseudotyped viruses expressing gp160 derived from 4E10 escape virus strains from patients NAB01 and NAB02 were measured on TZM-bl cells. ICs (IC50, clear bars; IC70, dashed bars; IC90, filled bars) of the wild-type virus (pre) and 4E10 escape mutants are depicted. The respective mutations in the core epitopes of the envelope clones are indicated.

FIG. 7.

Infectivity of neutralization escape viruses. The influence of 2G12, 2F5, and 4E10 resistance-conferring mutations in the envelope on the infectivity of the virus from patient NAB01 was studied. The env genes of the mock-treated control (black) and 2G12 escape (NAB01 2G12 II wk 10, gray), 2F5 escape (NAB01 2F5 I wk 8, dashed), and 4E10 escape (NAB01 4E10 II wk 11, white) mutants of the isolate from patient NAB01 were introduced into the replication-competent NL4-3 backbone TN6 NL and compared for infectivity on three-way-stimulated PBMC. The infectivity (TCID50/pg p24) of the viral stocks is displayed. Means of two independent experiments, each performed with pools of three independent donor PBMC, are shown.

FIG. 8.

Stability of escape mutations without (w/o) antibody (Ab) pressure. Patient NAB01 isolates selected to escape 2G12 (NAB01 2G12 II wk 10) (left panel), 2F5 (NAB01 2F5 I wk 8) (middle panel), and 4E10 (NAB01 4E10 II wk 11) (right panel) were subjected to long-term in vitro culture in the absence of antibody pressure. Inhibitory doses of the respective antibodies for virus strains derived during these long-term cultures are depicted (IC50, clear bars; IC70, dashed bars; IC90, filled bars). wt, wild type.

FIG. 9.

Clonal analysis of the MPER after long-term culture. MPERs of 2F5-resistant (NAB01 2F5 I wk 8) and 4E10-resistant (NAB01 4E10 II wk 11) virus strains obtained after long-term culture without (w/o) antibody (Ab) pressure. Sequences of the respective antibody epitopes were analyzed to detect whether reversion to the respective wild-type (wt) sequences (ELDKWA and SWFDIT for the epitopes of 2F5 and 4E10, respectively) occurs.